Again a 7854 power supply beast to be repaired...


unclebanjoman
 

I got this 7854 to repair.
Main symptom:it blows the 4A main fuse. After several tests and measurements, the culprit was found in one of the two leveling capacitors (C16), 950uF 200V with a serious leakage.
I replaced both capacitors C16 and C17 for safety. The SMPS chassis sits on my workbench, fully extracted from the mainframe.

I put a dummy load across +15 and -15 V (one resistor 33 ohm 10W), insert a jumper on P72 on LV board (thermal switch) to avoid a false FAULT sense condition and power up the SMPS.
The fuse does not blows. Good. The security indicator neon lamp is steady on.
I got the somewhat expected tick-tick from the SMPS. Damn...
I already read the various topics on this forum regarding the infamous 7xxx SMPS and all the possible documentation and/or links about that.
So I proceed with measurements and investigation.

With my surprise, the tick-tick is not as usual (one tick/burst every 500 ms or so) but is very irregular
With my oscilloscope I see short random pulses followed by one, two sometimes even three or more bursts (pin 9 of U75, base drive control).
On pin 1 of U75 I see a signal that grows towards + 0.6V but in a very irregular manner (in correspondence with the very short pulses) and then reach 0.6 / 0.7V when (in a random manner) the normal burst begins.
At this point I am totally puzzled because due to this random behavior I cannot understand how the correct signals should be.

To be sure, I've disconnected completely the LV regulator board and I've put the dummy load directly across the +17/-17V.
I checked all the voltage levels on the low-voltage rectifier section: during the burst they reach more or less the nominal value, so there are no short circuits or lossy capacitors. The +108V are present.

The fault sense pin shows an irregular signal of 50mV or less so I don't think that's the cause.
The over-voltage stop seems O.K. in the sense that the emitter of Q45 is LESS positive than its base. The zener diode VR45 is correctly regulating to 120V.
I've checked with my analog VOM (ICE 680R) all the bjts. All of them seem OK. The programmable UJT Q30 is OK also.
Checked the power transistors Q34 and Q40: they are good.
Checked for leaky capacitors OK
Checked all the diodes OK

I've been working and studying on this beast for three days but with no results, except a big headache.

Any ideas and/or suggestions?

Max


Craig Sawyers <c.sawyers@...>
 

I've been working and studying on this beast for three days but with no results, except a big
headache.

Any ideas and/or suggestions?
Situation normal with the Tek switchers AKA "High efficiency power supply". There are three that just
sit here having defeated all attempts at fixing.

Craig


Steph L
 

Is another test-bed link needed re Fault Sense signal P83-4? Suggest trace the signal path. [It has a convoluted journey through Main Interface (sheet 3 where renamed "HV sense"), then through Z-axis/Focus PCB (sheet 13) and finally back to T14 power transformer.]

Steph


unclebanjoman
 

Steph,

I'm working with only boards A23 and A12. The low voltage regulator board A22 was disconnected from the rest of power supply, and there are no other connections. I pulled out connectors P52,P50 and P54 from A12 board to isolate the problem.
Doing so, the Fault Sense signal at pin 2 of U75 is unconnected, pulled to ground R66/C66 so I assume that this signal is not the culprit.
I've put a dummy load across +17 and -17V (connector P52) , 33 ohm resistor 10 W, I think that it's a reasonable load for the SMPS to start. Or not?
The only voltage really needed to the control IC is the 108V in the rectifier section of A12 board, together with the -17V. These voltages are somewhat normal during the burst mode but, as I stated in my previous message, the tick mode (burst) is very irregular and random, interspersed with single pulses. So no chance to understand what's is happening for now!
Still puzzled.

Max


Harvey White
 

On Sun, 21 Apr 2019 18:04:09 -0700, you wrote:

Steph,

I'm working with only boards A23 and A12. The low voltage regulator board A22 was disconnected from the rest of power supply, and there are no other connections. I pulled out connectors P52,P50 and P54 from A12 board to isolate the problem.
Doing so, the Fault Sense signal at pin 2 of U75 is unconnected, pulled to ground R66/C66 so I assume that this signal is not the culprit.
I've put a dummy load across +17 and -17V (connector P52) , 33 ohm resistor 10 W, I think that it's a reasonable load for the SMPS to start. Or not?
The only voltage really needed to the control IC is the 108V in the rectifier section of A12 board, together with the -17V. These voltages are somewhat normal during the burst mode but, as I stated in my previous message, the tick mode (burst) is very irregular and random, interspersed with single pulses. So no chance to understand what's is happening for now!
Still puzzled.
Ok, perhaps this is a reasonable interpretation.

1) the power supply ramps up voltages
2) if an overload (let's assume that, otherwise it's overvoltage) is
detected, then the power supply shuts down, causing the tick.
3) a good solid short to ground means a dependable state of ticking.

So if the ticks are irregular, then perhaps the fault is irregular.
Something is not shorted completely, but something may be overloading
the supply, just not all the time. You may be looking at something
that's drawing current intermittently.

So, assuming that you're getting something failing, then the question
is how to find it....

One thing you can do is to look at the voltage at each output of the
supply for an overvoltage. I'd look at the shutdown signal, something
somewhere may be shorting out, or trying to, and then the voltage at
each supply might be rising, or trying to rise. That this is a
repeating situation is to your advantage, since it's producing voltage
waveforms you may be able to measure.

Doing so will probably require you to analyze the exact mechanism of
the overvoltage/overcurrent and go from there.

I seem to remember some threads on a similar supply.

Harvey




Max



Dave Casey
 

When you replaced the bad capacitors, were the originals of the can style
with several ground tabs? Did you restore all those ground connections by
using caps of the same type, or otherwise restore the connections with
jumper wires? In some places in Tek power supplies, the capacitor can is
part of the circuit. Removing a can without replacing it (or installing
jumper wires instead) can result in sections of the board not being
grounded as intended.

Dave Casey

On Sun, Apr 21, 2019 at 10:11 AM unclebanjoman <mmazza@inrete.it> wrote:

I got this 7854 to repair.
Main symptom:it blows the 4A main fuse. After several tests and
measurements, the culprit was found in one of the two leveling capacitors
(C16), 950uF 200V with a serious leakage.
I replaced both capacitors C16 and C17 for safety. The SMPS chassis sits
on my workbench, fully extracted from the mainframe.

I put a dummy load across +15 and -15 V (one resistor 33 ohm 10W), insert
a jumper on P72 on LV board (thermal switch) to avoid a false FAULT sense
condition and power up the SMPS.
The fuse does not blows. Good. The security indicator neon lamp is steady
on.
I got the somewhat expected tick-tick from the SMPS. Damn...
I already read the various topics on this forum regarding the infamous
7xxx SMPS and all the possible documentation and/or links about that.
So I proceed with measurements and investigation.

With my surprise, the tick-tick is not as usual (one tick/burst every 500
ms or so) but is very irregular
With my oscilloscope I see short random pulses followed by one, two
sometimes even three or more bursts (pin 9 of U75, base drive control).
On pin 1 of U75 I see a signal that grows towards + 0.6V but in a very
irregular manner (in correspondence with the very short pulses) and then
reach 0.6 / 0.7V when (in a random manner) the normal burst begins.
At this point I am totally puzzled because due to this random behavior I
cannot understand how the correct signals should be.

To be sure, I've disconnected completely the LV regulator board and I've
put the dummy load directly across the +17/-17V.
I checked all the voltage levels on the low-voltage rectifier section:
during the burst they reach more or less the nominal value, so there are no
short circuits or lossy capacitors. The +108V are present.

The fault sense pin shows an irregular signal of 50mV or less so I don't
think that's the cause.
The over-voltage stop seems O.K. in the sense that the emitter of Q45 is
LESS positive than its base. The zener diode VR45 is correctly regulating
to 120V.
I've checked with my analog VOM (ICE 680R) all the bjts. All of them seem
OK. The programmable UJT Q30 is OK also.
Checked the power transistors Q34 and Q40: they are good.
Checked for leaky capacitors OK
Checked all the diodes OK

I've been working and studying on this beast for three days but with no
results, except a big headache.

Any ideas and/or suggestions?

Max




Göran Krusell
 

Hi, I worked on this PS in a 7904 some years ago. There are two overload conditions, over/under voltage and over current. Either shall turn the supply off. However, I think there is a third condition that will turn the regulator off which is not quite obvious. If you have arcing in the multiplier for example you will get sudden and random load changes and thus disturbances that will destroy the phase of the signal coupled to pins 10 and 11. The phase detector will go nuts and the regulator will turn off, I think.

Can you hear a hissing sound, if so it could indicate arcing?

Meanwhile, there are a couple of tests that you can do but it takes some courage. Short pin 2 (FAULT SENSE) and then pin 13 (ISENSE) to ground. If the behaviour remains the same then shut down is not triggered through these ports.

Carry on
Göran


unclebanjoman
 

Following the Goran suggestion, I've shorted to ground pin 2 (FAULT SENSE).

The behavior was the same: random burst interspersed with single pulses.
Probing with an oscilloscope the pin 13 (Current Sense) it shows a constant voltage of about +300 mV with random single peaks reaching +400 mV and sometimes some pulses of 20 ms duration with amplitude of +400 mV too.

I remember you all the actual scenario:
I'm working with only boards A23 and A12 connected together.
The low voltage regulator board A22 was disconnected from the rest of power supply, and there are no other connections (I pulled out connectors P52,P50 and P54 from A12 board to better isolate the problem).
I've put a dummy load across +17 and -17V (connector P52) , 33 ohm resistor 10 W, I think it's a reasonable load for the SMPS to start.

Max


Göran Krusell
 

In the archive there is a document scanned by Zenith, a specification 155-0067-02, pdf. For pin 13 there shall be two conditions:
Logic "0": 0.0 - 0.05 V
Logic "1": 0.85 - 10 V
Well, if you dare you can short pin 13 to ground...

Göran


Albert Otten
 

Hi Max,

From my own experiments with a 7704A SMPS I remember that it needed a much heavier load for proper operation. I tried with loads on the HV outputs and also with loads on the raw rectified outputs (LV disconnected). Other members (including Göran) found similar results. Dennis in his recent 7104/7854 .docx file uses 70 W total, distributed over the LV outputs. So why do you think about 35 W is enough?
I guess you also disconnected the supply to the HV unit? (Some other readers might not know that the HV unit is a separate unit here, not buried deep inside the smps box.)

It has been noted that the documentation of the control IC is not very trustworthy.

[BTW your 10W resistor will be glad that the smps is not working continuously!]

Albert


I remember you all the actual scenario:
I'm working with only boards A23 and A12 connected together.
The low voltage regulator board A22 was disconnected from the rest of power
supply, and there are no other connections (I pulled out connectors P52,P50
and P54 from A12 board to better isolate the problem).
I've put a dummy load across +17 and -17V (connector P52) , 33 ohm resistor 10
W, I think it's a reasonable load for the SMPS to start.

Max


Göran Krusell
 

In the archive there is another document "7k_PS_currents" describing typical loads, also scanned by Zenith. In my case the total load was 50W. I connected a number of power resistors to the various LV power supply outputs. It worked ok.
Göran


unclebanjoman
 

To Göran:
OK, shorted pin 13 to ground. Still same result: the SMPS refuses to starts and ticks herratically.
I've checked the signal across A and K of Q30 (the PUT): it fires up every 20 msec (I live in Italy) but there are some spurious pulses in between.
The pin 9 shows a steady +0.6V with sparse negative pulses going to approx 0 V.

To Albert:
I really don't know how a dummy load must be connected to raw outputs! Some people told me to load +17V was sufficient, some others the +108, now you and others say that the load MUST be distributed across the raw outputs! I'm really confused now...
And yes, I've disconnected the supply to HV unit obviously... since the SMPS block sits on my workbench, alone.
Some times ago (several months ago I suppose), a friend of mine fixed a 7904 SMPS; he told me that as a load he had connected a 20 ohm resistor on the +15 (regulated output) and nothing else and the smps worked ...


Albert Otten
 

Hi Max,

Well, I didn't say that the load MUST be distributed. For me reasons to do so were that I could use available 5 W and 10 W resistors and that I wasn't always sure about the maximum current an output could deliver. Your 10 W resistor will survive start attempts, but when an error has to be detected which tends to occur after several minutes then you have a problem.
I am really surprised that about 11 W was sufficient. IIRC in a 7904 the HV circuits are less easily to disconnect so maybe these formed a load also. At this moment I don't remember all the details, but I think 11 W would not even be enough to maintain the 7.5 zener voltage inside th IC. Anyway, over 50 W was my experience.

Albert


To Albert:
I really don't know how a dummy load must be connected to raw outputs! Some
people told me to load +17V was sufficient, some others the +108, now you and
others say that the load MUST be distributed across the raw outputs! I'm
really confused now...
And yes, I've disconnected the supply to HV unit obviously... since the SMPS
block sits on my workbench, alone.
Some times ago (several months ago I suppose), a friend of mine fixed a 7904
SMPS; he told me that as a load he had connected a 20 ohm resistor on the +15
(regulated output) and nothing else and the smps worked ...


Göran Krusell
 

You need a load of about 50W, possibly equally shared between positive and negative outputs. Either on the raw voltages or after the regulator board. You need to check all caps and diodes on the rectifier board, if you have not done this before. You can check diodes by making differential measurements across the diode. I found two bad power diodes this way.
Göran


Göran Krusell
 

More. You can check diodes by making differential measurements across the diode. I found two bad power diodes this way. You can do this measurement while the unit is in tic mode. I found the diodes conducting in both directions!
Göran


unclebanjoman
 

O.K Albert,
. I've found theDennis' documents you mentioned on how to build a complete dummy supply load for the 7104/7854 SMPS.
I will buy the necessary resistors... :-/

BUT the document only mentions the connections to be made (using the dummy load) towards P82.
IMHO the connections to P83 that performs the voltage sensing should also be wired in an appropriate manner.
Otherwise the low voltage regulator (board A22, schematic 16) cannot work!

Thanks Göran for your suggestion regarding diode checking.
I only checked them using my analog VOM Supertester ICE 680R in ohmic mode and no defects was revealed.
I will try to measure them using the oscilloscope and differential probe.

Max Mazza (from near Venice, Italy)


Albert Otten
 


I've found theDennis' documents you mentioned on how to build a complete dummy supply load for the 7104/7854 SMPS.
I will buy the necessary resistors... :-/

BUT the document only mentions the connections to be made (using the dummy load) towards P82.
IMHO the connections to P83 that performs the voltage sensing should also be wired in an appropriate manner.
Otherwise the low voltage regulator (board A22, schematic 16) cannot work!
Hi Max,
I checked that P83 can be left disconnected, like in many previous versions of power supplies. At the LV board there are resistors R76, R96 and R132 for this between the sense lines and the voltage output lines.

Are you using a variac? I have seen strange things when I cranked up the variac untile the spms starts. Later on I applied normal line voltage immediately.

Albert


unclebanjoman
 

No variac, I'm using normal line voltage (230 V).

Max


GeorgeP
 

I have two of these to be repaired. A third one working. Following. What a machine is 7854! Thanks for the thread.


unclebanjoman
 

Welcome in the hell of 7854 SMPS repairing GeorgeP!
Your contribution is more than welcome!
Can you post some relevant images of the signal present at various test points of board A12 and A23 when in tick (burst) mode?

Max